클래스/형상 함수 변환법을 이용한 3차원 프로펠러 팬의 형상 파라미터화

Abstract

In this thesis, we used CST method for representing 3-D propeller fan. Class/shape function could be used for optimization by using limited number of design parameters, contrary to other existing parameterization method using aerodynamic shape. In addition, class/shape function transformation method provides analytic and well behaved geometric shape. Cross sectional shape of propeller fan is expressed by using CST method and generated cross section is distributed through shape function for distribution. However, contrary to an existing parameterization method of aircraft wing, as a distinctive parameter that is present in propeller fan only exists, parameterization will be performed by considering such element additionally. By using CST method, point is extracted through MATLAB and actual 3-D shape could be implemented through NX. And in order to confirm how well propeller fan shape being generated through CST method could express original shape, point fitting was progressed by using information about propeller fan that is used for outdoor A/C ventilator of Samsung Electronics. Point fitting was performed by using EA (Evolutionary algorithm) of PIAnO (Process Integration automation and Optimization) that is an optimal design program. In 1st time analysis of EA for point fitting, app. 24 hours was required by computer (specification: CPU-i7 3.20GHz RAM 12GB) based on popsize of 130, generation frequency of 2000 times and it was confirmed that the biggest error between measured shape and CST shape was represented as 3.0882~3.4233mm (0.908~1.0%). However, an accurate value could be deduced by reducing this error if number of design parameter should be increased through function degree. Through this result, it was judged that there would be no problem even if an optimal design should be performed by defining initial shape through CST method on behalf of actual model of 3-D propeller fan.